Видео

What shortcuts are you using in phpstorm with vim?

Great video 👍 ! but what will i miss if i used DB::transaction ?

How about increased memory Usage? Because most developers don’t think about optimisation.

This is such an interesting video, I'm working with a bunch of business logic that has executes commands on remote Linux servers (its a backup tool, so it takes files from ubuntu servers and sticks them on S3 buckets etc), and mocking it has been a frustrating experience. It doesn't inspire confidence when your script breaks and your tests are just happily passing like nothing happened due to it being so heavily mocked. 🤦 I'll have to try using fake objects in a small portion and see how I get on. Thank you Mateus!

Great video Mateus, thanks for sharing

Great Doctrine idea :)

Great video! Well done!

nice! This is somehow similar to MikroORM persist

I wish that Taylor left you some feedback on why not merging this instead of just closing with the auto message, guess it was because it's pretty big change to the Eloquent-way of thinking. Great idea for a package nevertheless, can see myself pulling this if the right scenario comes up.

Wow. This is awesome! It makes so much sense regarding consistency and I think I'm going to be spending some time over the next week bringing it into my latest side project. Thanks for doing this!

This is awesome. Thanks for teaching me. :)

Great idea, Mateus! I'll try it. I think this is worth a PR in Laravel 😉

Nice, I had no idea about that "persist" method. Thx!

push() exists on models already and is intended to save the model and relationships. (still a good idea to call it inside of a transaction)

Thanks for this! I really struggle with testing code that interacts with third party services, I'm not the best at mocking so its such a headache.

The whole point is that you get to abstract away where and how the data is stored. That's why it's a Repository, not a Service etc. I'm using Repositories in Laravel to pull in data from CSVs, APIs, MS-SQL, and MySQL, and the code to read and write that data is identical, regardless of the source / storage and retrieval mechanism. I can iterate over different repositories, calling the same methods on each of them, for example. In a sense, the Interface of the Repository is the most important part; the key abstraction layer.

Great video! How would you use docker to containerize something like this?

As much you did other videos on the topic, I think you can minimize referring to the other video now and then, you can just explain the topic and we move on

This was a great video. Just what I needed to solve a problem I've been having. Thanks Mateus!

Great video! One thing worth mention is that the model will likely change when we're switching from a sync vs. async approach. For instance, the Order approach, if it was sync, we could create the Order model only after the credit card processing succeeded (or maybe just commit the transaction then). But if you're going *async*, you'd probably need to create a *pending Order* that would be sent to the ProcessOrder job, which would run in background, and would update the Order and notify the customer afterward... something like that. So, going from sync to async (and vice versa, I guess) causes domain changes too.

Great video, go ahead Mateus!

Wow. It is first time I see something like that. It is unique. Everybody talks about microservices and blah blah. But before switching to microservices, very few people talk. Good for you!

great vid, people need to realize this is the kind of stuff that matters when talking about scale, not just the req/sec your js framework can handle.

It's really awesome concept

$command->customer->id, $order->payment->customer etc violate encapsulation

really great video. thank you a billion. keep it going

Great

I read about Aggregates from DDD a long time ago, and didn't quite figured out them for Laravel ORM (I think it is mostly a thing for Data Mapper) but this helped me to make it sink properly, thank you, still not quite sure why you have `Ticket` and `OrderTicket` Models, and why you pass a `Ticket` to the `addTicket` method and it creates an `OrderTicket` is `Ticket` kind of a ValueObject or a DTO?

Love this video, want more content like this in the future. Thanks.

how to modify x-app-layout could u make a video

You probably won't have a ton of line items, however, what if you're dealing with a domain model that could have a lot of "line items"? If you keep pushing into the array, you'll run out of memory. But this is not a "domain" problem. How to solve this?

Great video thanks Mateus!

Love your channel and videos, I am learning something new with every one. And I always think about how this could benefit my code base. Thank you!

rojla rak ,9adekech ksouh ya m3allam !

Just use the data mapper pattern lol, active record is such an anti pattern.

Great video. Thanks for sharing these thoughts. With things like OrderLines - which actually only exist within the order - what do you think of the approach of denormalizing the data structure and not storing it in a separate database table but as embedded entities? This would possibly solve some "problems" with persistence (and of course bring new challenges^^). Keep up the good work. I really enjoying your content.

oh god this is perfect. i'm a django developer and i've suffered these issues a lot. i tried learning DDD but couldn't get anything because i was thinking domain models = orm models. but it all made sense in the first 5 minutes of your video. thank uuuuu

This kind of more intermediate/advanced content is dearly missed on youtube! Great explanation and example, albeit simplified does illustrate the issue and possible solution in an eloquent way.. i mean elegant 😂

Concepts are too advanced for me but great video.

Something just doesnt sit right with me when business-related method is mixed with Eloquent "config" method on a Model. Especially when you have the Order model, you __can__ still call save() and all that. The way I see it, there are 3 methods in the save order "flow": start(), addTicket(), place(). Yet they are defined amidst a dozens of Eloquent specific methods: relationship, accessor and mutators, casts,... On top of the inherited Model methods: saves(), delete(), update(),... If someone would onboard 5 years from now, how would he know that these 3 methods are special and belong in a "flow", and that he should not call save() directly? How would he know to organize additional such methods? If we are talking about Aggregates, we are obviously implementing DDD, or at least DDD-like patterns, which separate business logic from infrastructure implementation. I think it best we dont put Domain Models (which handle business logic) and ActiveRecord models (which is atrongly coupled to db) inside each other

One thing this also does is put off any database transactions until the end. I've found that if you wrap everything into one database transaction at first, you end up stuffing a bunch of stuff in there without thinking about it, and those things (e.g., network calls, memory allocations for instances, etc.) all incur a cost. This can lead to long running transactions in your database which can lead to all manner of issues. This is an elegant pattern. Thanks!

I confess I'm just watching your videos because your voice 🤣 Another great vídeo !

Great job, Mateus! Any plans on doing something related to Rails anytime soon? Big fan over here, keep it up!

Hi Mateus. Could you help me understand the difference between failed jobs in Horizon and the `failed_jobs` table? Horizon seems to have its own concept of failed jobs that it stores in Redis. Is retrying a job using `queue:retry` the same as clicking "Retry" on a job in the Horizon UI? Are the 2 supposed to be kept in sync and work together? And is `queue:retry all` the right way to retry all failed jobs when I'm using Horizon?

i have problem in production on ce2

Also - do you use a factory pattern like managers (or drivers) and could you show a bit more about that as well as a Null object pattern?

Nice! I really dig this way of working. Question: Are you always using "Providers" to integrate with 3rd party solutions or do you also use "Services"? Sometimes I think it's hard to see the difference 😅 Also, would love to see a video of how to test each of the implementations. Would you in those cases use mock with http fake to fake the expected output of the api?

Fakes are indispensable in my opinion and I've started moving away from mocks whenever I have situations like this. Also, just watching your video made me think that a fake of an implementation is a sister of the Null Object Pattern. While used differently, they're of the same species at least. In addition to this, this is a great way to Mr. Miyagi someone into understanding the strategy pattern!

Keep them coming! Super informative!

I tried to listen but the cat made too much noise.